And yes, I think there are both valuable secrets and embarrassingly-bad implementations in the source.
So any conspiracy about Windows 95 having secret hooks that killed DR-DOS or slowed down WordPerfect in favor of MS Word would have to accept as a given that IBM was in on the gig then and was still remaining silent to this day even as it openly competes with MicroSoft by backing Linux.
Thus, even if Windows 95 isn't in the Library of Congress, IBM has seen most if not all of that code already during their earlier OS/2 co-development.
I mean, the whole secret conspiracy thing makes great tech-talk around the water coolers and at the local bars, but getting around the U.S. copyright situation as well as somehow keeping everyone at IBM silent about secret hooks and such is a bit past the believability point.